Professional Documents
Culture Documents
Wood
377
1. INTRODUCTION
Electro-optic effects belong to the larger class of nonlinear
optics phenomena.1 In nonlinear optics, the coefficients
are either calculated from a microscopic theory or experimentally determined.2 Because a coherent microscopic
theory of ferroelectrics is lacking, and the only one available at present is a phenomenological (free-energy)
theory,318 the only way to gain extra information about
some of the electro-optic coefficients is to determine them
experimentally.
The electro-optic effect is the change of the index ellipsoid of a crystal with the application of an electric field.
For crystals with inversion symmetry (12 point groups),
the linear electro-optic (Pockels) effect does not exist; the
quadratic electro-optic (Kerr) effect is the dominant one.
In crystals lacking inversion symmetry (noncentrosymmetric crystals), both the linear and the quadratic effects
are simultaneously possible, with the Pockels effect being
the dominant one. If one designs an experiment configured in such a way that the individual products r ij E j
0 with i, j 1, 2, 3, then the Pockels effect does not
contribute to the results, and the Kerr effect provides the
primary contribution to the data. In that case the r ij are
the first-order electro-optic coefficients, and the E j are the
components of the externally applied electric field. The
Kerr coefficients can be extracted by use of the theory presented in Section 2.
To test our method, we measured the R 12 and R 13 Kerr
coefficients for barium titanate in the tetragonal state,
where both the Pockels effect and the Kerr effect are
present. At temperatures above 120 C (Curie temperature), barium titanate is in the paraelectric state, which is
cubic with point group symmetry m3m. The quadratic
electro-optic effect, which is the main one for this phase,
1084-7529/2005/020377-08$15.00
2. THEORY
This section is divided into two parts. In the first part we
demonstrate how contributions from the Pockels effect
were avoided. In the second part we deduce the system
of nonlinear equations from which the off-diagonal Kerr
coefficients are extracted numerically from the experimental data.
2005 Optical Society of America
378
For the most general case (a biaxial crystal), the equation of the initial index ellipsoid (initial indicatrix) at a
zero electric field is given by
x2
y2
n 12
z2
n 22
n 32
1.
(1)
(2)
y2
n 2 2
z2
n 3 2
2yz
n 4 2
2xz
n 5 2
2xy
n 6 2
(3)
y 2
n 2 2
z 2
n 3 2
tan 2 5
tan 2 6
1.
(4)
2r 4j E j
1/n 2 1/n 3 2
2
2r 5j E j
1/n 1 2 1/n 3 2
2r 6j E j
1/n 1 1/n 2 2
2
(5)
(6)
(7)
22 2;
23 , 32 4;
33 3;
13 , 31 5;
12 , 21 6.
With this notation, the Pockels and Kerr expressions connecting the new refractive indexes (n ), the initial ones
(n), and the three components (E 1 , E 2 , E 3 ) of the applied
electric field in the (x, y, z) system of coordinates can be
explicitly written as
1
n i
1
n i2
n 1 2
1
n 2 2
r ij E j R i kl E k E l ,
(8)
n 3 2
1
n 12
1
n 22
1
n 32
r 1j E j R 1 kl E k E l ,
(9)
r 2j E j R 2 kl E k E l ,
(10)
r 3j E j R 3 kl E k E l .
(11)
tan 2 5
2r 51E 1
1/n 2o
1/n 2e
379
(13)
1/n 1 2
1/n 2 2
1/n 3 2
1/n 4 2
1/n 5 2
1/n 6 2
1/n 2o
1/n 2o
1/n 2e
0
0
0
r 13
r 13
r 33
r 51
r 51
R 12
R 13
R 12
R 11
R 13
R 13
R 13
R 33
R 44
R 44
R 66
(15)
A z1 t z A z0 ,
(16)
E1
E2
E3
R 11
E 12
E 22
E 32
E 2E 3
E 1E 3
E 1E 2
(14)
(12)
380
ty
tz
1 r 2y exp i y
1 r 2y exp 2i y
1 r 2y 2 cos y i 1 r 4y sin y
1
r 4y
2r 2y
cos 2 y
(17)
T y t y 2
exp 2i z
1 r z2 2 cos z i 1 r z4 sin z
1 r z4 2r z2 cos 2 z
T yz t y t z* t *
y tz 2
rz
2
1
2
T yz sin 2 ,
(25)
T yz sin 2 ,
ny 1
ny 1
nz 1
nz 1
T z t z
1 r 2y 2
1 r 4y 2r 2y cos 2 y
1 r z2 2
(18)
(26)
1 r z4 2r z2 cos 2 z
(27)
(28)
1 r 2y 1 r z2 1 r 2y r z2 cos r 2y r z2 cos y z
1 r 4y 2r 2y cos 2 y 1 r z4 2r z2 cos 2 z
These transmission coefficients take into account the effect of multiple reflections at the two faces of the crystal.
The r y , r z are the Fresnel reflection coefficients at normal
incidence; they are related to the indexes of refraction n y ,
n z in the following way:
ry
I z A 02 T z cos2 T y sin2
with
1 r z2 exp i z
r z2
I y A 02 T y cos2 T z sin2
(19)
(20)
(29)
z y .
(30)
The amplitude modulated signal oscillates with a frequency directly proportional to the phase difference .
On top of this signal, there are many small interference
fringes generated by the multiple reflections at the two
faces of the crystal; their frequencies are directly proportional to the phases y and z . To calculate the net effect of these rapid fringes on the amplitude of the modulated signal, we average the transmissivities T y , T z and
the mixed term T yz over a 2 interval for each of the two
phases y and z . Considering the as an independent variable, we obtain
T y
1
2 2
T y y d y d z
1 r 2y
1 r 2y
,
(31)
y 2 / n y l x ,
(21)
z 2 / n z l x ,
(22)
(23)
A z A z1 cos A y1 sin .
(24)
T z
1
2 2
T z z d y d z
1 r z2
1 r z2
,
(32)
T yz
1
2 2
T yz y , z d y d z
1 r 2y r z2 cos
1 r 2y 1 r z2
(33)
I y A 02 T y cos2 T z sin2
I z A 02 T z cos2 T y sin2
1
2
T yz sin 2 ,
(34)
1
2
T yz sin 2 .
(35)
381
(36)
I yr
I y
I y0
(37)
I zr
I z
I z0
(38)
where T y0 , T z0 , and T ( yz ) 0 are the transmission coefficients and the mixed transmission term in the initial
state of the crystal, before the application of an electric
field. For normalization purposes, so that I yr 1 and
I zr 1 when E 1 0 and s x 0, we chose the denominators in Eqs. (37) and (38) to have a form analogous to
their corresponding numerators when E 1 and s x are zero.
3. EXPERIMENT
The experimental design is schematically shown in Fig. 3;
all data were collected at a temperature of 23.5 C. The
sample was an unclamped 5 mm 5 mm 1 mm, twoface-polished, (100)-oriented, single crystal of barium titanate. It was mounted with the spontaneous polarization Ps in the z direction. A pair of electrodes 3 mm in
diameter were created by sputtering a 20-nm thin layer of
gold on the 5 mm 5 mm opposite faces of the sample.
The connection between these electrodes and the electri-
382
Fig. 4. Plot of I y , I z , and I m versus time for the first experimental trial. The applied electric field (E 1 ) across the BaTiO3
sample was switched in a steplike fashion approximately every
50 s for 500 s. The values of the electric voltage taken in chronological order are 0, 400, 400, 0, 600, 600, 0, 200, 200,
0 V.
Fig. 5. Plot of I y , I z , and I m versus time for the second experimental trial. The applied electric field (E 1 ) across the BaTiO3
sample was switched in a steplike fashion approximately every
50 s for 500 s. The values of the electric voltage taken in chronological order are 0, 400, 400, 0, 600, 600, 600, 0, 200,
0 V. This time the 400 V voltage drop on the sample was kept
on for only 20 s.
383
, I zr
and Off-Diagonal Kerr Coefficients R 12 , R 13 for
Table 1. Orthogonally Polarized Intensities I yr
Externally Applied Electric Fields E 1 , First Experimental Trial
I yr
E 1 (103 V/m)
400
400
600
600
200
200
Trial 1 Average
0.80
0.78
0.66
0.71
0.88
0.85
R 12(1017 m2 /V2 )
I zr
0.02
0.02
0.04
0.02
0.01
0.02
1.23
1.20
3.67
1.15
1.11
1.11
0.03
0.05
0.23
0.03
0.01
0.02
3.51
3.59
1.44
3.70
3.43
3.42
3.5
0.11
0.11
0.08
0.11
0.06
0.11
0.3
R 13(1017 m2 /V2 )
7.82
7.92
7.58
6.26
8.19
8.21
7.9
0.13
0.31
0.49
0.19
0.07
0.13
0.5
, I zr
and Off-Diagonal Kerr Coefficients R 12 , R 13 for
Table 2. Orthogonally Polarized Intensities I yr
Externally Applied Electric Fields E 1 , Second Experimental Trial
I yr
E 1 (103 V/m)
400
400
600
600
600
200
Trial 2 Average
0.70
0.61
0.57
0.47
0.61
0.87
R 12(1017 m2 /V2 )
I zr
0.03
0.02
0.02
0.04
0.02
0.02
1.37
1.50
1.38
1.67
1.31
1.13
5. CONCLUSION
We have shown a general method for measuring the offdiagonal coefficients of the quadratic electro-optic (Kerr)
tensor and verified its usefulness by determining the two
off-diagonal
Kerr
coefficients,
R 12 3.5 0.3
1017 m2 /V2 and R 13 8.0 0.7 1017 m2 /V2 for
a (100), 5 mm 5 mm 1 mm, single crystal of BaTiO3 .
The method relies on selecting an experimental configuration in which the individual products r ij E j 0 with i,
j 1, 2, 3. Unlike other methods,3645 our method treats
multiple reflections at the two faces of the sample, light
transmission through the crystal, and electrostriction.
Our method can be used to measure off-diagonal Kerr coefficients on 18 of the 20 crystal point groups in which the
Pockels effect and the Kerr effect coexist.
0.03
0.02
0.03
0.06
0.03
0.02
3.46
3.43
3.53
3.47
3.54
3.41
3.4
0.17
0.11
0.11
0.23
0.11
0.11
0.2
R 13(1017 m2 /V2 )
8.07
8.16
7.75
8.02
7.73
8.26
8.1
0.19
0.13
0.19
0.37
0.19
0.13
0.4
The values for these two quadratic electro-optic coefficients were numerically extracted from the abovementioned set of nonlinear equations [Eqs. (37) and (38)].
The uncertainties result primarily from nonlinear dielectric behavior with the electric field, which is a common
phenomenon in every ferroelectric crystal.
The fact that the two intensities I y and I z can be associated with the symbols 1 and 0 of Boolean algebra might
make it possible for the working principle of this light
modulator to find some applications in the improvement
or design of new optical binary devices that are so needed
in the important technological field of information processing optical devices (optical computers).
Using this technique, one can monitor the changes in
the optical coefficients of crystals under different external
conditions, i.e., electric or magnetic field and pressure and
temperature variations. These measurements can provide the physics community with a new window into the
process of symmetry change occurring in crystals immediately after they are subjected to an external stress.
Furthermore, the knowledge gained from observing the
time evolution of crystal behavior could also be used to
improve the state of theoretical modeling associated with
the dynamics of phase transitions in some of these crystals.
As the two graphs (Figs. 4 and 5) show, a curious behavior of the light intensities as functions of time
(t), I zr
(t)) was detected every time the applied elec(I yr
tric field was changed. The authors have not seen these
effects discussed in the scientific literature to date. We
suspect that some of these light transients are caused by
the change in the indexes of refraction caused by a certain
relaxation in the crystals polarization once the electric
stress has been temporarely removed. A more concrete
and detailed study of the connection between this tran-
384
22.
ACKNOWLEDGMENTS
The authors acknowledge support from the Texas Higher
Education Coordinating Board Advanced Research Program. We also thank C. L. Chen and X. H. Chen for depositing the gold electrodes on the sample.
L. T. Woods e-mail address is ltwood@uh.edu.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.